J. Mirapeix

Fast algorithm for spectral processing with application to on-line welding quality assurance

A new technique is presented in this paper for the analysis of welding process emission spectra to accurately estimate in real-time the plasma electronic temperature. The estimation of the electronic temperature of the plasma, through the analysis of the emission lines from multiple atomic species, may be used to monitor possible perturbations during the welding process. Unlike traditional techniques, which usually involve peak fitting to Voigt functions using the Levenberg-Marquardt recursive method, sub-pixel algorithms are used to more accurately estimate the central wavelength of the peaks. Three different sub-pixel algorithms will be analysed and compared, and it will be shown that the LPO (linear phase operator) sub-pixel algorithm is a better solution within the proposed system. Experimental tests during TIG-welding using a fibre optic to capture the arc light, together with a low cost CCD-based spectrometer, show that some typical defects associated with perturbations in the electron temperature can be easily detected and identified with this technique. A typical processing time for multiple peak analysis is less than 20 ms running on a conventional PC.

Spectroscopic analysis of the plasma continuum radiation for on-line arc-welding defect detection

When plasma optical spectroscopy is applied to on-line welding quality monitoring, the plasma electronic temperature is commonly selected as the spectroscopic parameter to determine. However, several processing stages have to be considered in this case, including plasma emission line identification, which is significantly costly in terms of computational performance. In this paper, the wavelength associated with the maximum intensity of the plasma background emission is proposed as the new monitoring signal, as it will be demonstrated that there is a clear correlation between this parameter and the welding quality. The resulting processing scheme is clearly simpler, and experimental and field tests will prove the feasibility of the proposed technique.

Overcoming Nonlocal Effects and Brillouin Threshold Limitations in Brillouin Optical Time-Domain Sensors

We demonstrate, for the first time to our knowledge, a Brillouin optical time-domain analysis (BOTDA) sensor that is able to operate with a probe power larger than the Brillouin threshold of the deployed sensing fiber and that is free from detrimental nonlocal effects. The technique is based on a dual-probe-sideband setup in which an optical frequency modulation of the probe waves along the fiber is introduced. This makes the optical frequency of the Brillouin interactions induced by each probe wave on the pump vary along the fiber so that two broadband Brillouin gain and loss spectra that perfectly compensate are created. As a consequence, the pulse spectral components remain undistorted, avoiding nonlocal effects. Therefore, very large probe power can be injected, which improves the signal-to-noise ratio (SNR) in detection for long-range BOTDA. Moreover, the probe power can even exceed the Brillouin threshold limit due to its frequency modulation, which reduces the effective amplification of spontaneous Brillouin scattering in the fiber. Experiments demonstrate the technique in a 50-km sensing link in which 8 dBm of probe power is injected.

Data Processing Method Applying Principal Component Analysis and Spectral Angle Mapper for Imaging Spectroscopic Sensors

A data processing method to classify hyperspectral images from an imaging spectroscopic sensor is evaluated. Each image contains the whole diffuse reflectance spectra of the analyzed material for all the spatial positions along a specific line of vision. The implemented linear algorithm comes to solve real time constrains typical of industrial systems. This processing method is composed of two blocks: data compression is performed by means of principal component analysis (PCA) and the spectral interpretation algorithm for classification is the spectral angle mapper (SAM). This strategy, applying PCA and SAM, has been successfully tested for online raw material sorting in the tobacco industry, where the desired raw material (tobacco leaves) should be discriminated from other unwanted spurious materials, such as plastic, cardboard, leather, feathers, candy paper, etc. Hyperspectral images are recorded by a sensor consisting of a monochromatic camera and a passive prism-grating-prism device. Performance results are compared with a spectral interpretation algorithm based on artificial neural networks (ANN).

Defect Detection in Arc-Welding Processes by Means of the Line-to-Continuum Method and Feature Selection

Plasma optical spectroscopy is widely employed in on-line welding diagnostics. The determination of the plasma electron temperature, which is typically selected as the output monitoring parameter, implies the identification of the atomic emission lines. As a consequence, additional processing stages are required with a direct impact on the real time performance of the technique. The line-to-continuum method is a feasible alternative spectroscopic approach and it is particularly interesting in terms of its computational efficiency. However, the monitoring signal highly depends on the chosen emission line. In this paper, a feature selection methodology is proposed to solve the uncertainty regarding the selection of the optimum spectral band, which allows the employment of the line-to-continuum method for on-line welding diagnostics. Field test results have been conducted to demonstrate the feasibility of the solution.

Sensor for the Detection of Protective Coating Traces on Boron Steel With Aluminium–Silicon Covering by Means of Laser-Induced Breakdown Spectroscopy and Support Vector Machines

Welding processes are one of the most widespread industrial activities, and their quality control, in both online and offline methods, is an important area of research. In the particular process of laser welding of boron steel with aluminium-silicon covering in the automotive industry, one problem is the presence of residual traces from the protective antioxidant coating, an aluminium-silicon alloy, which can result in a significant reduction of the welding seam strength. This work proposes a sensor system based on a laser induced breakdown spectroscopy (LIBS) setup to detect and discriminate aluminium residues in the welding area without destroying the sample before the welding procedure. A spectral algorithm based on support vector machines (SVMs) is used as a classifier to automatically identify areas with aluminum presence.

Plasma spectroscopy analysis technique based on optimization algorithms and spectral synthesis for arc-welding quality assurance

A new plasma spectroscopy analysis technique based on the generation of synthetic spectra by means of optimization processes is presented in this paper. The technique has been developed for its application in arc-welding quality assurance. The new approach has been checked through several experimental tests, yielding results in reasonably good agreement with the ones offered by the traditional spectroscopic analysis technique.

Proposal of Brillouin optical frequency-domain reflectometry (BOFDR)

We demonstrate a Brillouin optical frequency-domain reflectometry (BOFDR) technique, which can measure the strain and/or temperature along an optical fiber with one-end access, by detecting the spontaneous Brillouin scattering from a sinusoidally modulated pump light. Compared to the Brillouin optical frequency-domain analysis (BOFDA), we show that BOFDR measurements are free from the distorting components related to acoustic wave modulation, thus simplifying data processing.

Use of the Plasma Spectrum RMS Signal for Arc-Welding Diagnostics

A new spectroscopic parameter is used in this paper for on-line arc-welding quality monitoring. Plasma spectroscopy applied to welding diagnostics has typically relied on the estimation of the plasma electronic temperature, as there is a known correlation between this parameter and the quality of the seams. However, the practical use of this parameter gives rise to some uncertainties that could provoke ambiguous results. For an efficient on-line welding monitoring system, it is essential to prevent the appearance of false alarms, as well as to detect all the possible defects. In this regard, we propose the use of the root mean square signal of the welding plasma spectra, as this parameter will be proven to exhibit a good correlation with the quality of the resulting seams. Results corresponding to several arc-welding field tests performed on Inconel and titanium specimens will be discussed and compared to non-destructive evaluation techniques.

Arc-Welding Spectroscopic Monitoring based on Feature Selection and Neural Networks

A new spectral processing technique designed for application in the on-line detection and classification of arc-welding defects is presented in this paper. A non-invasive fiber sensor embedded within a TIG torch collects the plasma radiation originated during the welding process. The spectral information is then processed in two consecutive stages. A compression algorithm is first applied to the data, allowing real-time analysis. The selected spectral bands are then used to feed a classification algorithm, which will be demonstrated to provide an efficient weld defect detection and classification. The results obtained with the proposed technique are compared to a similar processing scheme presented in previous works, giving rise to an improvement in the performance of the monitoring system.